Software Engineering (9th Edition) [Hardcover]

I had the displeasure of using this book for a graduate level course on Software Engineering. We did not use the web site or the online learning tools associated with the book. The professor did use the author's supplied slides.

Pros:
* Very nice UML diagrams.
* Might get recycled into toilet paper.

Cons:
* Very repetitive. I lost track of how many times a legacy system was defined. This made reading this book extremely boring.
* Despite the 2011 copyright date, the material is dated. Computers are still single core, smart phones aren't on the scene, and Sun still owns Java (cue Oracle lawyers).
* The back of the book claims it has been updated with new material on open source development. That new material consists of a few paragraphs on the legal issues of incorporating open source into a traditional project. There is nothing on developing software for open source.
* The power point slides that accompany the book have problems with the graphics starting about chapter 5. The image quality of the embedded diagrams takes a nose dive and the images are barely readable.
* The topics covered seemed very shallow. I'm not sure you'll get much more out of this book then you would reading through wikipedia articles.
* The author has a habit of using acronyms without defining them. COTS (Commercial Off The Shelf) was used for several chapters before it was spelled out.
* Electronic version (Kindle) has random spaces removed (as noted by other reviewers)
* The project schedule charts presented are Gantt charts. It seems the author has never heard of Gantt and just refers to them as bar charts.
* Some diagrams are mislabeled.
* There was at least one sample XML file that was used for a few problems. The XML was broken making it that much harder to figure out the problems.

If you are a professor looking for a book on this topic, please spare your students and find another book.
If you want almost the whole book for free, download the slides. Chapter 4 alone is 82 slides (29 pages in the book).

The only thing I'd like to get out of this book (besides my money back) is how the diagrams were created. I haven't found anything that comes close to the ones in the book.

This book has such a poor rating in Amazon, and most of the reviewers feel the book lacks examples. It IS true.
However, I believe the author is NOT trying to provide the readers every detail about everything he is talking about. If I would teach a SWE course, the book would take me 2-4 semesters to cover. Just think about it, I can write a series of books talking about testing alone, about architecture alone, about XP in Agile alone, about......... This is not the book that give you the whole detail. However, why do so many professors choose this book as their textbooks(the 9th edition at least shows something)? Because it DOES provide a great overview about the key problems in SWE and most of the popular solutions to those problems. Also, it is the professor's discussion in the lectures and the labs that are more important to students who really wanna learn more about SWE. You MUST do projects, coding, putting those theories to practice to know that SWE really is. Even if the author put thousands of questions on the back of the book won't help you learn SWE.

My advice:
Get the big picture from the book, get the detail from the professors and other good resources. Learn what SWE really is through projects.

To those who wanna learn SWE by yourself:
ok......I have to say that is really tough. If you are in a company, you may don't have a lot of time, otherwise you may don't have project opportunity. As this book contains a lot of information while not that much details, it will drive you crazy sometime. So, this book only worth 2.5 stars for you :< the 2.5 stars are for you to grasp the big picture.

Any book in its ninth edition has already proven its place to its audience - Sommerville's text has earned its reputation as a standard in the field. This book's twenty-six chapters cover four major topics: introductory material, dependability and security, advanced software techniques, and project management. The first section include basic life cycle phases of requirements, modeling, architecture, implementation, testing, and ongoing support, along with smattering of the social processes in which these activities take place.

Advanced topics, the third section, include modern topics such as aspect-oriented programming (AOP), component systems, and embedded applications. AOP seems to be moving in from the fringes of software development (Aspect-Oriented Programming with the e Verification Language gives lots of good reasons for that, despite being tangential to normal software). Still, object orientation dominates current practice by a wide margin and gets only minor mention, so I find the emphasis misplaced. Likewise, the embedded section under-represents the 99% of all processors that don't run Windows or Linux, i.e., the ones in your microwave, digital watch, CD player, car air bags, sewing machine and cell phone. Still, mentioning it at all puts this ahead of many comparable texts.

So does the second section, on dependability and security. As computers become more pervasive and take on more life-critical applications, these issues only grow in importance. As with other sections, however, topics that represent many people's entire careers get only 20 or 30 pages here. A book of this breadth necessarily sacrifices depth, however, so I can't complain too loudly. Likewise, the section on project management issues could be a book of its own, like Braude's Software Engineering: An Object-Oriented Perspective.

Real discussion of this book's content would take more space than I have here, and would soon get into matters that depend on specific applications or on personal opinion and experience. I found one pervasive weakness in this book, however, a weakness shared by every other software development and engineering text I've seen. Standards and regulatory requirements get only scant mention. In the real world, software design and implementation happens in a complex environment of interlocking standards that affect nearly every aspect of software design and implementation. I applaud Sommerville for mentioning Sorbanes-Oxley (despite the book's international audience). Still, the hundreds of applicable standards from the IEEE, ECMA, ISO, ITU, NIST, and dozens of other organizations warrant a chapter of their own. I look forward to seeing that chapter, in this book if not in some other.

- wiredweird

PS: I'm wrong a lot, but that cover photo could credibly be MIT's Stata center under construction. If so, in restrospect now that it's completed, that wouldn't be the most uplifting choice of engineering images.